LeadIQ – Autonomous CALL-E Voice Lead Qualification & Sales

🚀 The Problem

Sales reps waste 65% of their day on cold calls and manual qualification. When a hot lead picks up, slow CRM updates delay AE outreach by 24–48 hours — killing deals.

💡 What LeadIQ Does

LeadIQ turns CALL-E into a 24/7 autonomous phone outreach team:

1. Dials prospect lists with E.164 normalization & idempotency protection
2. Qualifies via natural voice conversation — budget, timeline, decision-maker, pain points
3. Analyzes transcripts with Gemini 2.0 Flash to extract structured buying signals
4. Routes hot leads (fit >80%) to Senior AEs via Slack with a 2-hour SLA

🔧 How We Built It

• CALL-E REST API (/v1/calls) for outbound voice with webhook callbacks
• Gemini 2.0 Flash for transcript qualification → structured JSON
• Multi-tier fallback: Gemini → Ollama → deterministic heuristics (never fails offline)
• Flask REST backend with SQLite persistence + Google Sheets sync
• Glassmorphic dashboard — live pipeline, lead intelligence drawer, CSV export

🏆 Accomplishments

• 100% autonomous pipeline: CSV → voice call → AI qualification → AE routed in <30s
• 15/15 tests passing on Python 3.9+
• Real CALL-E API integration with idempotency, webhook handler, result_schema
• Interactive Lead Intelligence Drawer with verbatim call quotes & pain point tags

📚 What We Learned

CALL-E's result_schema enables structured data extraction directly from voice. Gemini at temperature 0.2 produces highly consistent qualification JSON. Idempotency keys are essential at scale.

🔮 What's Next

CRM integrations (Salesforce, HubSpot), inbound call handling, A/B script testing, and a SaaS multi-tenant version.

Built With

• calle
• flask
• gemini
• llm
• python
• vector